iSgz-J ly ['Davis. 
ice was melting away so as to allow an unequal settling of tlie 
drumlin above it. Such crevices are entirely too exceptional to 
be used in corroboration of so extremely specialized a condition as 
an under layer of ice, even if they could not be explained by the 
unequal settling of less solid drift masses on which, as well as on 
ice, the drumlin may have been originally accumulated. 
It is postulated that drundins were rapidly accumulated, and 
hence that some sijecial process of origin such as is now advanced 
is necessary for their explanation ; but here again, the postulate is 
not universally accepted ; and even if it were, the ]>ossibility of 
drumlins -being gathered up with comparative rapidity from a 
drift encumbered district in a re-advance of the ice is not ex- 
cluded. As a whole, the theory seems to me to be based on 
special interpretations of phenomena that might be differently 
inter])reted ; and I am therefore unable to assert that drundins 
were actually formed in this {)articular way, however possible the 
way may be. 
Wiien we come to the verification of Mr. Upham's theory by 
the success of its consequences in meeting various facts of obser- 
vation, we find two classes of difficulty: first, the variety of 
special consequences that follows from this theory and from no 
other is small, and hence demonstration by frequent accordance, 
that is, by excessively jirobable correctness, is not reached ; second , 
certain special consequences are not clearly confirmed by the 
facts. I will consider only one other of these, in addition to 
the examples already mentioned. 
In applying the theorj'^ to special cases, such as College Hill, a 
large drumlin a few miles northwest of Boston, I find serious dif- 
ficiilties at every step. College Hill is not more than two miles 
from the northern boundary of the blue slates of the Boston Basin, 
yet it contains plentiful fragments of these slates. Now, if the 
drumlin is formed by the process advocated by Mr. Upham, it 
must be supposed that the slate fragments were first gathered 
from ledges at the bottom of the ice ; then raised to a considerable 
height in the ice mass by tlie inferred differential velocities of 
flow ; next transformed from their englacial state to the condition 
of superglacial drift by surface melting, wherel>y the height at 
which many of them stood within the ice must have been reduced ; 
then washed downward and in a general way southward over the 
melting ice in the process of surface concentration ; again 
